While mail surveys offer an efficient and cost effective means of gathering large amounts of student-related data in a relatively short time, surveys characterized by low response rates may have a biasing effect on information. To examine this bias, a study was undertaken of the characteristics associated with samples from three graduate follow-up surveys conducted at CCP between 1989 and 1991. Demographic characteristics, entering academic abilities, academic engagement, and persistence were examined for survey respondents, voluntary nonrespondents who received the survey and chose not to participate, and involuntary nonrespondents who did not receive the questionnaire due to an incorrect address. The most consistent group differences were associated with academic performance and persistence. Respondents earned higher grade point averages (GPA's) and persisted for a longer time than nonrespondents. Voluntary nonrespondents were less academically successful and persisted for longer periods than involuntary nonrespondents. Results indicate that survey results may be skewed since the more positive a student's academic experience, the more likely she/he is to respond. The following strategies are suggested to reduce the number of nonresponses: (1) send a short, professional introductory letter and a simple, logical questionnaire; (2) include a self-addressed stamped envelope; (3) personalize the letter; (4) send a second survey to nonrespondents; and (5) remind students of the importance of making address changes. Contains 19 references. (KP)